Paper 4, Section II, H

Optimization | Part IB, 2021

(a) Consider the linear program

P: maximise over x0,cTx subject to Ax=b\begin{aligned} P: \quad \text { maximise over } x \geqslant 0, & c^{T} x \\ \text { subject to } & A x=b \end{aligned}

where ARm×n,cRnA \in \mathbb{R}^{m \times n}, c \in \mathbb{R}^{n} and bRmb \in \mathbb{R}^{m}. What is meant by a basic feasible solution?

(b) Prove that if PP has a finite maximum, then there exists a solution that is a basic feasible solution.

(c) Now consider the optimization problem

Q: maximise over x0,cTxdTxQ: \quad \text { maximise over } x \geqslant 0, \quad \frac{c^{T} x}{d^{T} x}

subject to Ax=bA x=b,

dTx>0,d^{T} x>0,

where matrix AA and vectors c,bc, b are as in the problem PP, and dRnd \in \mathbb{R}^{n}. Suppose there exists a solution xx^{*} to QQ. Further consider the linear program

R: maximise over y0,t0,cTyAy=btdTy=1\begin{aligned} R: \quad \text { maximise over } y \geqslant 0, t \geqslant 0, & c^{T} y \\ & A y=b t \\ & d^{T} y=1 \end{aligned}

(i) Suppose di>0d_{i}>0 for all i=1,,ni=1, \ldots, n. Show that the maximum of RR is finite and at least as large as that of QQ.

(ii) Suppose, in addition to the condition in part (i), that the entries of AA are strictly positive. Show that the maximum of RR is equal to that of QQ.

(iii) Let B\mathcal{B} be the set of basic feasible solutions of the linear program PP. Assuming the conditions in parts (i) and (ii) above, show that

cTxdTx=maxxBcTxdTx\frac{c^{T} x^{*}}{d^{T} x^{*}}=\max _{x \in \mathcal{B}} \frac{c^{T} x}{d^{T} x}

[Hint: Argue that if (y,t)(y, t) is in the set A\mathcal{A} of basic feasible solutions to RR, then y/tB.]y / t \in \mathcal{B} .]

Typos? Please submit corrections to this page on GitHub.